-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Le 28/06/2015 16:01, clipka a écrit :
> Am 28.06.2015 um 14:19 schrieb Le_Forgeron:
> 
>> The same code seems to be already in 3.1g
>> 
> ...
>> The hypercomplex of povray seems to match the Davenport's type
> 
> Take a closer look at the initialization of the iteration, in the 
> 3-parameter overload of the
> HypercomplexFunctionFractalRules::Iterate function. Still happy
> with it?
> 

That code (in master branch, because in stable it is in
Iteration_HCompl_Func inside hcmplx.cpp) seems ok in regard to julia
computation:

 v_(n+1) <-- Function(v_n) + c, repeat for n from 0 to max_iteration-1
unless it goes too far from center (that's a classical optimisation:
once the point get far enough from center, it is known that it will
not converge again)

Thanks for the light of the computation stack (and it confirm that in
3.1g, if you had more than one julia with a different limit on the
number of interaction ( *max_iteration* count), you were in trouble.
Deep trouble of out of range memory write & read.

from http://www.povray.org/documentation/view/3.6.1/280/ the
hypercomplex julia is due to FractInt .

The initialisation of the iteration (n=0) seems fine, but for one
thing: The Sz[0]/IterStack[Z][0] is getting IPoint[Y] instead of
IPoint[Z] .

That's a big typo, if you consider the result, but a single letter in
thousand of line of code :-)


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iJwEAQEIAAYFAlWQMdIACgkQhKAm8mTpkW3lZQQAlBxc0pojVF1bl0n8gzqH6FUU
cb64Pr37ItZelG2SyrOPwq8M1uIWH405aiq5dp2U+E6hCtqwPGztgjn+y3t40pzh
FkoAhHyj2AoTeO1SCAH+narMAmqr+swWGnnMzz1UvH5fGnu/3JkEwE9pQtG5sKeM
5JAZg11MRQF2PIa9pv4=
=3V6j
-----END PGP SIGNATURE-----

Post a reply to this message